Yellow Pages US Scraper

Type a trade and a city and get the businesses trading there, as rows you can sell to: name, phone, email, full address, website, services, opening hours, coordinates and social profiles. Every business's own page is opened to collect the full record, not just the search result. Adverts are left out unless you ask for them, duplicates are removed before you are charged, and a business with no phone can be dropped before it ever reaches your list.

🔍 What does Yellow Pages US Scraper do?

You name the trades and the places, and it returns one row per unique business.

There are two ways to say what to collect, and they work together in one run:

• Trades and places. Put plumbers into searchTerms and Austin, TX into locations. Every trade is searched in every place, so three trades across four cities is twelve searches in one run.
• Pages you paste. Put a Yellow Pages search or category page into urls and it is walked exactly as given. Mix them with the trades above.

The easiest way to try it is to leave everything alone and click Start. The defaults are a working job.

Four things are worth understanding before you run it, because they are the whole product.

Every business's own page is read, not just the search results. A Yellow Pages search result carries a name, a phone and a link. The email, the services, the payment methods, the coordinates and the social profiles are only on the business's own page, so this Actor opens that page for every business it returns. That is what turns a directory listing into something you can run a campaign from. Turn includeDetails off if you only want the search fields and want them fast.

Adverts are not results, and here they are not mixed in. Yellow Pages puts paid placements above and among the real listings. They are usually national call centres with an 800 number rather than a business in the city you searched, and every tool that returns them mixed in is selling you a lead list you then have to clean. This one leaves them out by default. Turn includeAds on if you want them, and they arrive marked isAd so you can still tell them apart.

You pay per unique business, after deduplication. The same business appears on several pages of one search and across two searches that overlap, which is exactly what happens when you scan a metro area by trade. Before anything is delivered or charged, businesses are deduplicated on Yellow Pages' own id, so the number on your bill is the number of usable rows in your dataset. Duplicates are counted, reported, and cost you nothing.

A business you filtered out is never charged. requirePhone, requireWebsite and minReviews are read before the row is counted, so a lead list of only callable businesses costs you only the callable ones.

⬇️ Input

Field	Type	Default	What it does
searchTerms	array	["plumbers"]	Trades to search for, one per line
locations	array	["Austin, TX"]	Cities or ZIP codes. Required when you give a trade
urls	array	empty	Yellow Pages pages to walk as given
maxResults	integer	200	Unique businesses to return in total. This is what you pay for
maxPagesPerSearch	integer	10	How deep to go into each search. About 30 businesses a page
includeDetails	boolean	true	Open each business's own page for email, services, coordinates and socials
includeAds	boolean	false	Include paid placements, marked isAd
requirePhone	boolean	false	Drop businesses with no phone
requireWebsite	boolean	false	Drop businesses with no website
minReviews	integer	empty	Drop businesses under this many reviews

Limits worth knowing: up to 50 searches in one run (trades multiplied by places), a search term up to 120 characters, and up to 20,000 businesses. Each is refused with a message saying which limit it was, before anything is collected or charged.

A trade always needs a place. Without one, Yellow Pages answers for wherever the run happened to exit, which on a rotating address pool is a different city every run. The run refuses rather than handing you a dataset that looks right and is not.

{
  "searchTerms": ["plumbers", "hvac contractors"],
  "locations": ["Austin, TX", "Round Rock, TX"],
  "maxResults": 500,
  "requirePhone": true
}

⬆️ Output

One row per unique business. The fields below the line come from the business's own page and are collected when includeDetails is on, which it is by default.

	Field	What it holds
🆔	ypId	Yellow Pages' own id, which is what the run deduplicates on
🏢	name	The business name as listed
☎️	phone	The listed number, in the grouping a caller dials
📍	street	Street address. Null on a business that serves an area
📍	city, state, zip	The address, split so you can filter and sort on it
🔗	website	The website the listing links to, where it has one
🏷️	categories	The trades Yellow Pages files the business under
⭐	reviewCount	How many reviews the listing carries. Most listings have none: 19 of 200 on the measured run
⭐	rating	The star rating, 1 to 5 in half steps. Null where there are no reviews
🕒	openingHours	Opening hours as published, such as Mo-Fr 09:00-17:00. Null where the listing gives none
📅	yearsInBusiness	Years trading, where the listing claims any
📢	isAd	Whether this is a paid placement rather than a result
🔗	ypUrl	The business page on Yellow Pages, for checking a row by hand
📝	snippet	The description line under the listing
🧭	foundVia	The search page the row came from
🕐	scrapedAt	When the row was collected, ISO 8601 UTC
✉️	email	The business email, where the business published one
🔧	services	The individual services the business lists, split into separate entries
📝	description	The business's own description of itself, useful for personalising an approach
💳	paymentMethods	What the business accepts, such as visa, cash, amex
🗺️	areaServed	The area the business says it covers, often wider than its address
📌	latitude, longitude	Coordinates, for mapping and radius filtering
📅	yearEstablished	The year the business says it was founded
🗣️	languages	Languages the business says it speaks besides English
🖼️	logo	The business photo or logo
👍	facebook, twitter, instagram, linkedin, youtube	Social profiles, each in its own column
✅	detailScraped	Whether the business's own page was read. See the note below

detailScraped matters when a column is empty. A blank email on a row where detailScraped is true means Yellow Pages holds no email for that business. A blank on a row where it is false means the page could not be read and the answer is unknown. Every other tool leaves you unable to tell those apart.

How full are the detail fields?

They are as full as the business made them, and no fuller. Measured on 60 plumbers in Austin, TX on 29 August 2026, with every detail page read successfully:

Field	Businesses that published one
name, phone, website, categories, logo	100%
areaServed, city, state, zip	97%
latitude and longitude, openingHours	80%
description	77%
paymentMethods	65%
services	55%
yearEstablished	45%
email	25%
rating and reviewCount	25%
social profiles	2%

These are printed rather than hidden because a lead list is bought on what it actually contains. Around a quarter of US Yellow Pages businesses publish an email; the rest publish a website instead, and website is on effectively all of them. If you need an address for every business, run the domains from this dataset through an email finder.

💡 Why scrape Yellow Pages?

Territory lead lists. A trade and a list of towns gives you every business in that market with a phone number, which is the input to a cold calling list or a direct mail run.

Enrichment input. Turn requireWebsite on and the output is a domain list, which feeds straight into an email finder or a tech stack check. A quarter of businesses publish an email here directly; the domain column covers the rest.

Market sizing. How many roofers trade in a metro area, how many carry reviews, and how long they have been going, is three columns of this dataset and a pivot table.

Competitor mapping. Categories, services and review counts across a city say who the established firms are and what they sell before you price against them.

Map and radius work. Coordinates come with the row, so a dataset drops straight into a map or a drive time filter without a geocoding step or a geocoding bill.

💰 How much does it cost?

You pay $0.0008 per unique business delivered, and nothing else beyond Apify's own start charge. A thousand businesses is $0.80 on the free plan. Every paid Apify plan pays less through the Bronze, Silver, Gold, Platinum and Diamond discount tiers, down to $0.65 per thousand on Diamond.

Four things about that price are worth being explicit about:

• A duplicate costs nothing. A business found on four pages is one row and one charge.
• A filtered business costs nothing. An advert, or a business without the phone you required, is dropped before it is counted.
• A run that finds nothing pays the start charge and no more. A search with no results, a refused address and a blocked page are all delivered as a sentence rather than a bill.
• The detail pass is included in that price, not charged on top. Opening every business's own page is roughly thirty times the requests of reading search results alone. It is inside the per business rate rather than a second charge, so the number on your bill is still the number of rows in your dataset.
• Yellow Pages sits behind a Cloudflare challenge, which no ordinary request gets past whatever headers it sends. This Actor goes through an unblocking proxy that solves it, so the page arrives as plain HTML and no browser is paid for. That is what keeps the full record affordable at this price.

maxResults is a hard cap, so the most any run can cost is maxResults times the rate, plus the start charge.

On the run above, 200 businesses cost $0.16 on the free plan and $0.13 on Diamond. A thousand is $0.80 and $0.65.

⏱️ How long does a run take?

Reading every business's own page is what takes the time. A search result page hands over about thirty businesses in one request; the detail pass is one request per business, so a full run is roughly thirty times the requests of a listing-only one.

Measured on 29 August 2026, with every detail page read successfully:

Businesses asked for	With full details	Listing fields only
60, one search	2 minutes 47 seconds	23 seconds
200, four searches	7 minutes 43 seconds	74 seconds
1,000, five trades across four metros	26 minutes 14 seconds	10 minutes 10 seconds

Memory peaked at 80MB on the 1,000 business run, inside the 256MB this Actor asks for. Throughput settles at about 0.65 pages a second and does not improve by asking for more at once, because the unblocking proxy is the queue rather than this Actor.

If you want speed over depth, turn includeDetails off. You get the search result fields, in the times in the right hand column, at the same price per business.

Set the run timeout to suit the size of the ask. The run stops cleanly inside whatever timeout it is given and tells you it did, rather than being killed with the work lost. Measured: a 120 business ask given a 90 second timeout finished in 69 seconds, delivered the 51 businesses it had, and said so. Allow about 30 minutes per thousand with details on, or about 15 minutes with them off. The default is 3600 seconds, which covers the largest run in the table above. You are charged per business delivered rather than per minute, so a generous timeout costs you nothing.

A search that runs out of new businesses stops on its own rather than walking empty pages, so a thin trade in a small town finishes quickly and cheaply.

🔌 Integrations

The output schema is declared, so this plugs into Make, Zapier, n8n, Google Sheets, Airtable and Slack through Apify's own integrations without anyone writing a parser first. Point a webhook at the run and a finished dataset can land in a CRM on its own.

🔄 API usage

curl -X POST "https://api.apify.com/v2/acts/spookyweb~yellow-pages-scraper/runs?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"searchTerms": ["plumbers"], "locations": ["Austin, TX"], "maxResults": 200, "requirePhone": true}'

❓ FAQ

Do I need a Yellow Pages account or an API key?

No. There is nothing to register and no key to manage.

Why do I need to give a location?

Because Yellow Pages will silently pick one for you based on where the run exits, and on a rotating address pool that is a different city each time. A dataset built that way looks correct and is not, which is worse than an error.

Why are adverts excluded by default?

Because an advert is not a lead. The paid slots are usually national call centres rather than businesses in the city you searched, and a list you have to clean before you can call it is not the thing you bought.

How complete are the fields?

Every business carries an id, a name, a phone, a website and categories. The rest depends on what the business itself published, and the measured coverage for each field is in the table further up rather than left for you to discover after paying.

Why is the email column empty on some rows?

Because about three quarters of US Yellow Pages businesses never published one. Where a row has detailScraped set to true and no email, that is Yellow Pages' complete record of that business and no tool can return one from this source. Where detailScraped is false, the page could not be read and the answer is unknown.

Almost every business does publish a website. If you need an address for all of them, take the website column and run it through an email finder.

Can I skip the detail pages and just get the search results?

Yes. Set includeDetails to false. The run finishes several times faster and costs the same per business, but the rows carry no email, services, coordinates or social profiles.

How many businesses can I get from one search?

Yellow Pages serves about 30 a page and paginates deeply, but a given trade in a given town has a real floor. When a search runs out, the run says so in its status message and stops rather than walking empty pages at your expense.

Are the phone numbers formatted?

They are kept exactly as Yellow Pages published them, which is the grouping a person reads and dials. Reformatting them would lose the area code separation.

Does it get blocked?

Yellow Pages runs a Cloudflare challenge. This Actor goes through an unblocking proxy that solves it, and when a page is challenged anyway it retries on a fresh route rather than on the one that just failed. When a page cannot be read at all, the run says so and charges nothing for it.

Is scraping Yellow Pages legal?

This collects information published openly on the public web, with no login and no account, which is the same information a person browsing the site can see. It collects business contact details rather than personal data. You are responsible for how you use the output, including under GDPR, CAN-SPAM, TCPA and any local rules on unsolicited contact, and for honouring do-not-call registers before you dial a list.

Something is wrong or missing

Open an issue on the Issues tab. Layout changes on the source site are the usual cause and are usually a quick fix.

🙋 Feedback

Ideas, missing fields and bug reports all go on the Issues tab, and they get read.

🔎 You might also like

• Company Email Finder, to turn the websites in this output into work email addresses.
• Website Contact Scraper, to pull emails, phones and socials from the sites this finds.
• Company Data Enrichment, to add company details to a list of domains.